Evaluation of tumor ischemia in response to an indole-based vascular disrupting agent using BLI and (19)F MRI

Abstract

Vascular disrupting agents (VDAs) have been proposed as an effective broad spectrum approach to cancer therapy, by inducing ischemia leading to hypoxia and cell death. A novel VDA (OXi8007) was recently reported to show rapid acute selective shutdown of tumor vasculature based on color-Doppler ultrasound. We have now expanded investigations to noninvasively assess perfusion and hypoxiation of orthotopic human MDA-MB-231/luc breast tumor xenografts following the administration of OXi8007 based on dynamic bioluminescence imaging (BLI) and magnetic resonance imaging (MRI). BLI showed significantly lower signal four hours after the administration of OXi8007, which was very similar to the response to combretastatin A-4P (CA4P), but the effect lasted considerably longer, with the BLI signal remaining depressed at 72 hrs. Meanwhile, control tumors exhibited minimal change. Oximetry used (19)F MRI of the reporter molecule hexafluorobenzene and FREDOM (Fluorocarbon Relaxometry using Echo Planar Imaging for Dynamic Oxygen Mapping) to assess pO2 distributions during air and oxygen breathing. pO2 decreased significantly upon the administration of OXi8007 during oxygen breathing (from 122 ± 64 to 34 ± 20 Torr), with further decrease upon switching the gas to air (pO2 = 17 ± 9 Torr). pO2 maps indicated intra-tumor heterogeneity in response to OXi8007, though ultimately all tumor regions became hypoxic. Both BLI and FREDOM showed the efficacy of OXi8007. The pO2 changes measured by FREDOM may be crucial for future study of combined therapy.

Keywords: 19F MRI; CA4P; OXi8007; Vascular disrupting agents (VDAs); bioluminescence imaging (BLI); hexafluorobenzene; hypoxia; oximetry.

Figure 1

Molecular structure of OXi8007.

Figure 2

BLI monitoring of 231/luc orthotopic breast tumor response to IP administration of saline, CA4P or OXi8007. A. Maximum BLI signal was observed around 15 minutes post SC injection of luciferin at baseline, 4, 24, 48 and 72 hours post injection of saline (control), CA4P and OXi8007 in representative mice. B-D. Corresponding dynamic time courses of total flux obtained at the five time points for these three mice (B. Saline; C. CA4P; D. OXi8007).

Figure 3

Summary of BLI results for the treatment groups. (A) Relative maximum light flux and (B) integrated area under the light emission curve following treatment: saline (control, n = 4), CA4P (120 mg/kg; n = 4) and OXi8007 (350 mg/kg; n = 6). Error bars represent standard error.

Figure 4

Dynamic pO₂ changes assessed using FREDOM. A. pO₂ maps of a representative tumor during baseline air breathing; 20 mins after switch to O₂-breathing; 2 hours after administration of OXi8007 (350 mg/kg), while continuing to breathe O₂ and after final switch back to air (left to right). B. Histogram showing pO₂ distributions at the four stages. Mean pO₂ and hypoxic fractions (HF₅ and HF₁₀) were respectively: baseline air-breathing (three dynamics), 47 Torr, 4%, 5%; baseline O₂-breathing (last three dynamics before VDA administration), 228 Torr, 0.7%, 0.7%; O₂-breathing two hours post OXi8007 administration (last five dynamics with stable pO₂ measurements before switching gas back to air), 58 Torr, 5%, 7%; air-breathing post OXi8007 (last three dynamics with stable pO₂ measurements), 12 Torr, 29%, 40%.

Figure 5

Intra-tumoral heterogeneity of response to O₂ breathing and OXi8007. A. Two ROIs were identified as more and less responsive in the tumor presented in Figure 4. B. Mean pO₂ of each region revealing differential dynamic response. Error bar representing standard error is used to demonstrate the spread of pixel pO₂ values within ROIs.

Figure 6

Tumor oxygenation with respect to intervention. Population mean and standard deviation of tumor pO₂ (A) and hypoxic fraction, HF₅ (B) in response to gas intervention and OXi8007. **indicating p < 0.005, *indicating p < 0.05.